Global DNA methylation analysis in relapsed Pre-B cell acute lymphoblastic leukemia

Reliable biomarkers for relapsed acute lymphoblastic leukemia are scarce. Currently, minimal residual disease (MRD) is the best method in predicting a relapse event, but is invasive to the patient. In addition, high quantity amount of cells are needed for flow-MRD and PCR-MRD requires stable IG-TCR rearrangements. An alternative MRD strategy may involve DNA based technologies involving mass spectrometry that have the potential to monitor the progression of ALL. There are also many recurrent karyotypes that stratify patients into either low-standard risk or high-risk of relapsing. However, many patients with favorable karyotypes (e.g. highhyperdiploidy, ETV6/RUNX1) still experience one or more relapse events. DNA methylation has the potential to serve as a biomarker throughout the course of the disease in several aspects. DNA methylation associated microarrays have demonstrated the ability to stratify patients into their recurrent cytogenetic subtype. Further, others have identified CpG loci that have the potential to stratify patients at diagnosis that are at risk of relapsing. This dissertation identified differential DNA methylation between matched diagnosis and relapsed patients by creating a methylome profile (MIRA-seq) for each patient. We are the first to report global hypomethylation that occurs at relapse, predominately within retrotransposable elements. In addition, genes that exhibited 5' regulatory aberrant methylation from diagnosis to relapse were identified. Some of the genes harboring epigenetic lesions may be considered an "epidriver" of ALL. A "driver" mutation can be defined as a mutation that can directly or indirectly offer a Darwinian advantage in terms of growth for leukemic blasts. The "epi" component refers to the nature of the mutation as epigenetic in origin and in the context of this study is synonymous with deviant or aberrant methylation. Further, our MIRA-seq study in canine acute leukemia observed epigenetic lesions of epidriver genes that were also present at diagnosis and relapse in human ALL. Lastly, MIRA-seq profiles from ALL patients at diagnosis who either did or did not go on to relapse were compared against each other in order to identify individual CpGs using pyrosequencing that were associated with relapse (prognostic). These combined efforts were done with the foresight of identifying potential novel targets that exist as epidrivers of ALL or loci that hold prognostic power at diagnosis. This dissertation builds upon others who have previously identified relapse-associated biomarkers with implications of improved patient care and risk stratification.